2003-Blanco-caron-lesca-scipDeveloping Capabilities to Create Collective Intelligence

8/13/2019 2003-Blanco-caron-lesca-scipDeveloping Capabilities to Create Collective Intelligence

1/15

Journal

of

CompetitiveIntelligence

and

ManagementIntroductionEditors Note ....................... ...................... .................p. ii

Competitive Intelligence Field Research:Moving the Field Forward by Setting a ResearchAgenda. Usha Ganesh, Cynthia E. Miree, and John Prescott ..................... ...................... ... p. 1-12

Chronological and Categorized Bibliography of KeyCompetitive Intelligence Scholarship: Part 1

(1997- present). Paul Dishman, Craig Fleisherand Victor Knip .................... ...................... .. p. 13-79

Developing Capabilities to Create CollectiveIntelligence within Organizations.Sylvie Blanco, Marie-Laurence Caron-Fasanand Humbert Lesca .....................................p. 80-92

Seeing and Noticing: an Optical Perspective onCompetitive Intelligence.Michael L. Neugarten...............................p. 93-104

Volume 1, Number 1, Spring 2003
http://www.scip.org/jcim.asphttp://www.scip.org/jcim.asphttp://www.scip.org/jcim.asphttp://www.scip.org/jcim.asphttp://www.scip.org/jcim.asphttp://www.scip.org/jcim.asphttp://www.scip.org/jcim.asphttp://www.scip.org/jcim.asphttp://www.scip.org/jcim.asphttp://www.scip.org/jcim.asphttp://www.scip.org/jcim.asp


2/15

Journal of Competitive Intelligence and Managem

Journal of Competitive Intelligence and Management Volume 1 Number 1 Spring 2003

The Journal of Competitive Intelligence and Man-agement (JCIM) is a quarterly, international, blindrefereed journal edited under the auspices of theSociety of Competitive Intelligence Professionals(SCIP). JCIM is the premier voice of the CompetitiveIntelligence (CI) profession and the main venue forscholarly material covering all aspects of the CI andmanagement field. Its primary aim is to further thedevelopment and professionalization of CI and toencourage greater understanding of the manage-ment of competition by publishing original, highquality, scholarly material in an easily readable for-mat with an eye toward practical applications.

Journal of

CompetitiveIntelligenceand

Management Edited by Craig S. Fleisher ([email protected])and John E. Prescott ([email protected] )Editorial BoardDavid Blenkhorn, Wilfrid Laurier University

Ontario, CanadaPatrick Bryant, University of Missouri,Kansas City, USA

Jonathan Calof, University of Ottawa, CanadaAlessandro Comai, ESADE, Barcelona, SpainBlaise Cronin, Indiana University , Indiana, USAPaul Dishman, Brigham Young University , Utah, USAPat Gibbons, University College, Dublin, IrelandBen Gilad, Academy of CI, USA/IsraelChristopher Hall, Macquarie University , NSW, AustraliaWilliam Hutchinson, Edith Cowan University

WA, AustraliaPer Jenster, Copenhagen Business School, DenmarkKwangsoo Kim, Konkuk University , KoreaPaul Kinsinger, Thunderbird University , Arizona, USAQihao Miao, Shanghai Library , China

Jerry Miller, Simmons College, Massachusetts, USACynthia Miree, Oakland University , Michigan, USASusan Myburgh, University of South Australia, Australia

Juro Nakagawa, Tokyo-Keizai University , JapanEdna Reid, Nanyang Technology University , SingaporeHelen Rothberg, Marist College, New York, USALuiz Felipe Serpa, Universidade Catolica de Brasili

BrazilKathy Shelfer, Drexel University , Pennsylvania, USATom Tao, Lehigh University , Pennsylvania, USA

Joaquin Tena, University of Pompeu Fabra, Spain Jim Underwood, Dallas Baptist University , USAConor Vibert, Acadia University , Nova Scotia, CanadaSheila Wright, DeMontfort University , UK


3/1580

Journal of Competitive Intelligence and Manage


Developing Capabilities toCreate Collective Intelligencewithin Organizations

Sylvie BlancoGrenoble Graduate School of Business (ESC Grenoble)

Marie-Laurence Caron-FasanUniversity of Grenoble, FranceHumbert LescaUniversity of Grenoble, France

Executive Summary

Although Business Intelligence (BI) is perceivedas being more and more essential to the survival of organizations, its viability and effectiveness can bequestioned in terms of the inability of practitioners toexploit strategic information. As little work is avail-able on this practical issue, our objective is to fill thegap by developing a method for the creation of

collective intelligence on organizational environments.Using a qualitative methodology known as engineer-ing management research, we have attempted tofurther both practical and theoretical knowledge aboutBI. So far, we have completed four experiments withinorganizations. The theoretical as well as the practicalresults are encouraging. In this article we have at-tempted to present our approach in a way that may be of value to people interested in applying it them-selves.

Key Words:Business Intelligence, exploitation of strategic in-

formation, collective intelligence, engineering man-agement research

About the AuthorsMarie-Laurence Carson-Fasan is currently work-

ing as an assistant professor of Business Administra-

tion specializing in Management of Information Sys-tems at the Graduate Business School of University of Grenoble (France). She received her Ph.D. in Informa-tion Systems from Grenoble University in 1997. Herresearch interests include Management InformationSystems, Business Intelligence, Cognitive Process, andHuman Information Processing. She is a consultantspecializing in BI implementation. She has written 11papers in international and French [email protected]
http://[email protected]/http://[email protected]/


4/1581

Developing Capabilities to Create Collective Intelligence within Organi


Humbert Lesca is a professor of Business Admin-istration specializing in Management Information Sys-tems at the Graduate Business School (E.S.A.) of theUniversity of Grenoble (FRANCE). His main researchinterests include Management Information Systems,Business Intelligence, Organizational Learning, andHuman Information Processing. He directs a Ph.D.course in Management Information Systems. At theC.E.R.A.G., he manages a research team of eightcollaborators with whom he is involved in variousUniversity-Entreprise Partnerships. He has writtenfour books and more than 30 papers for French andinternational [email protected]

Sylvie Blanco is currently working as an assistantprofessor and researcher in Technological Manage-ment at the Graduate School of Business of Grenoble(ESC Grenoble). She obtained her Ph.D. in Manage-ment Information Systems from Grenoble Universityin 1998. Her research interests include Managementof Information, Business Intelligence, and Manage-ment of Technology. She has been working as a BIconsultant over 8 years. [email protected]

IntroductionBusiness Intelligence (BI) can be defined as the

information process through which companies pro-spectively monitor their environment in order tocreate opportunities and to reduce their uncertainty(Lesca, 1994). On the basis of this definition, threemain statements may be made about BI.

First, as a strategic decision support tool designedfor anticipation purposes, BI deals with non-routineand unique decisions. Bounded rationality and ap-

proximate reasoning are therefore unavoidable toprocess information which restricts possibilities forimplementing algorithm-based and expert systemsapproaches. More specifically, BI may be included inthe intelligence stage of organizational and individualdecision-making processes as formulated by Simon(1982). It therefore involves a stage for informationsearch, interpretation and vision building rather thanthe implementation of rational models.

Secondly, the forward-looking nature of BI im-plies a focus on anticipatory information - what

Ansoff (1975) called weak signals. Their main featuresare 1) no intrinsic relevance and 2) no possible defini-tion in advance (Feldman & March, 1981) as to eitherthe content or the source of information. These twofeatures make it difficult to process this kind of information and may lead people to ignore it. Finally,BI is an information gathering process that can belinked to an iterative learning process of which themain steps are described in Figure 1 below.

A major problem in the field BI lies in the confron-tation between researchers assertions on how toperform BI and practitioners difficulties and lack of ability in implementing it. The purpose of this articleis to shed light on one of these difficulties: the use of weak signals to identify potential threats and oppor-tunities and to heighten the understanding of theprobable future.

To do so, we shall first formulate explanations of these difficulties and use them to build up a concep-tual approach to help practitioners in situation. Weshall then describe the instrumentation and practicalimplementation of this approach within organiza-tions. In doing so, we shall underline the majormethodological issues so that our experience may bereproduced by researchers and practitioners. We havetried to make sure that the prerequisites and condi-tions of implementation are observed. Finally, wehave presented both practical and methodologicalcontributions.

Selectingfilt ering weak signa ls and adding

anticipating interpretation

Targetingidentifiying environmental

a ctor s and the me s to b emonitored .

Sharing

storing informationwithina common database

Exploitingtransforming weak signals

in to driv ing forces

Action

Tracking

identifying environmentalscanners and g iv ing t hemm ea ns to c ol lec t ex te rna l d at a

Figure 1: The BI Process
http://[email protected]/http://www.esc-grenoble.fr/http://www.esc-grenoble.fr/http://www.esc-grenoble.fr/http://[email protected]/


5/1582



Background and Purpose

Paradox between Theory and PracticeRegarding BI, the gap between theory and prac-

tice is wide. According to management theory, man-agers know and are capable of assessing changes intheir socio-economic environment to seize opportuni-ties as they arise. In practice, managers do not relyvery much on anticipatory information. Our work isgrounded in the gap between theory and practice.The following case illustrates something of the para-dox observed in companies:

Sometimes, my colleagues urge me to con-sider pieces of information they have justcaptured. The problem is that I am over-

Consequences for Their Analysis

Weak signals are related to potential future events that mayaffect the organization. They must forewarn managers earlyenough for reaction to be possible. Hence, each signal does nothave much significance in itself and is quite difficult to relate toimmediate decisions to be taken.

Weak signals are not numbers with records, or extrapolations.They are related to potential events that have not yet occured andmay never occur. Therefore, signals that alert to future events cannot consist of either quantitative or factual data.

Weak signals do not consist of certainties but of clues andtraces. They can be interpreted in different ways with no possibil-ity of identifying the right interpretation or they cannot beinterpreted at all.Therefore, they are not easily captured.

Weak signals are present in the form of fragments which have been patiently collected and gathered by various environmentalscanners. Taken separately, each fragment is insignificant andsuspect. Hence, significance can only be achieved by patientcross-checking. It is a gradual process.

Pieces of information may be picked up in any shape or form,such as snatches of conversation, electronic data, messages fromconferences and so on. . . . As they are not homogenous, theirexploitation is all the more difficult.

Nature of Weak Signals

Anticipatory

Qualitative

Ambiguous

Fragmentary

Of Various Presentations

Table 1Nature of Weak Signals and Consequences of their Use

whelmed by information and their pieces of information rarely fall into my immediate wor-ries. Hence, I leave them aside for the time being and when those pieces of informationare finally required, I can seldom retrieve them.It is often too late when I come to understandthat they were surely strategic pieces of infor-mation.

- Director of a medium-size company

According to observations collected during ourfield experiences, this situation occurs frequently incompanies. The problem is that relevant weak signalsare seldom exploited afterwards because they are nolonger accessible when they are needed and require


6/1583



crosschecking to gain significance. We believe that theintrinsic nature of weak signals (Ansoff, 1975; Feldman& March, 1981; Lesca 1992; and Table 1) largelyaccounts for the paradox mentioned above. Finally,our research leads us to conclude that methods arelacking to help managers process strategic informa-tion.

Assumptions, Inference and Purpose of this Paper

AssumptionsIt is probably because weak signals captured by

environmental scanners have no immediate use andno obvious significance that managers perceive theirexploitation as being difficult. They would prefer toreceive weak signals in an appropriate form justwhen needed. However, since no methods are avail-able, managers fail to exploit weak signals and there-fore BI viability is often questioned within organiza-tions because no anticipatory and action-orientedrepresentations of the environment are achieved.

InferenceIf our assumptions are correct, there should be

general acceptance from most managers of a methodto analyze pieces of information and produce mean-ing, both individually and collectively, even whenthere is no rush to solve a problem or to make adecision.

Purpose of this PaperThe aim of this paper is to try to reduce the gap

we have identified by designing a new BI method andimplement it within several organizations. Feedback

from the experiments is provided, and emphasis isgiven to the utility and the practicability of the method.

Draft of a Method to Produce CollectiveIntelligence from Weak Signals

According to Gorry and Scott-Morton (1971), afeature that is lacking in information systems is theirinability to develop models that reflect the way man-agers see their organization and their environment.

Understanding managers cognitive process is sup-posed to be an essential condition for the design of aneffective decision support system (Gorry & Scott-Morton, 1971; Rowe & Ziti, 2000). We agree with thispoint of view and take it as one of our assumptions.Consequently, we accept that progress in BI informa-tion handling could be achieved by relying on humancognitive process.

Many authors (Miller, 1956; Mintzberg, 1976;Goldhar, Bragaw & Schwarts, 1976; Taggart & Robey,1981) have tried to represent human cognitive pro-cesses. Two main ideas are emerging from their mod-els: 1) the regrouping of information and 2) thecreation of links between pieces of information. Weshall argue both ideas in the following section.

Regrouping Pieces of InformationMcKenney and Keen (1974) have proposed a model

to describe the way people structure information(either oral or visual) which has been captured intheir environment. They suggest the use of a regroup-ing process. But the way this regrouping is performedneeds to be specified.

Weber (1984) also analyzed the human cognitiveprocess related to regrouping. When faced with ambiguous situations, people try to build meaningfulrepresentations of their environment by placing piecesof information side by side and grouping them to-gether.

Again, the regrouping method has to be specified.Therefore, we underline the need to formulate assign-ment criteria enabling new pieces of information to beclassified easily into existing or new groups undereveryday pressure. These criteria must be usable, both individually and collectively, and explicitly com-municable to others in order to create collective intel-ligence. Two criteria have been mentioned in therelevant literature:

Similarity CriterionPieces of information can be grouped by similar-

ity. We try to connect similar information whether itexpresses the same idea or relates to the same theme.Kawakita-Jiro (in Hogarth, 1980) uses this criterion.Each piece of information is assigned to the groupwith which there is a link. Users then find they are


7/1584



dealing with small groups of information that can behandled separately.

According to Moles (1990), when somebody isconfronted with a sketchy set of ill-matched ele-ments , she/he tries to find some similarity betweenthese elements. When faced with a new piece of information, people try to find some similarity withan existing group (Conklin, 1987). Behling, Giffordand Tolliver (1980) have proposed a method to con-nect a piece of information with a group. Pieces of information are assigned to a group if they havecommon characteristics with this group, so that eachnew piece of information, whether perceived or re-ceived, can be analyzed. After this, the informationwill either be rejected or assigned to a group. In the box illustrating the regrouping of pieces of informa-tion, they are placed side by side in non-particularorder, at least to begin with (Weber, 1984).

Proximity CriterionThis criterion is less restrictive than the first, but

also more approximate. One way of assigning piecesof information to an existing group is to use theproximity criterion. Proximity means that informa-tion seems close to the theme it will be connected to.Recognition of a common characteristic is a proximitycriterion. Pieces of information may be quite differ-ent, but individuals can bring them closer by using acommon point of view. According to Moles (1990),this assignment can be done easily because it isnatural: when faced with a new piece of information,people include it in an appropriate existing group because they sense its proximity. This is a subjectiveprocess and is probably done when taking into ac-count a major individual preoccupation. But whenthey process the most familiar signals, people are notcareful enough about other signals that announcechanges (Barr, Stimpert & Huff, 1992). Hence there isa risk of biases arising from subjectivity. This risk isreduced if regrouping is done collectively.

Once this first step is completed, pieces of infor-mation are put together, side by side in no particularorder at least to begin with. But completing this stagedoes not in itself produce useful meaning. A secondstage is therefore necessary to show how pieces of information are connected.

Connecting Pieces of Information Withinand Between Groups

Another step can be completed by creating con-nections between pieces of information within andamong existing groups. Kawakita-Jiro (in Hogarth,1980) has proposed a creative technique to buildsignificant structures, from unconnected pieces of information at the moment when they are gathered.The basic idea of this technique involves intercon-necting pieces of information in each group. Theauthor specifies that each piece of information has to be compared to the others and matched in order toproduce a significant construct for users.

Lee and Lai (1991) have proposed seven types of

links: the logically implied link, the support link, thedenial link, the qualifying link, the presuppositionlink, the object to link and the answer link. We shalldeal with some of these links calling them by theirmost common name.

The Causality Link and the Influence LinkBougon, Weick and Binkhorst (1977) have shown,

through an observation study, that individuals usecausality connections to classify knowledge in theirminds. Information A is connected to information B if

A is the cause of B. In fact, causality is the mostcommon link used by authors such as Barr, Stimpertand Huff (1992), Narayanan and Fahey (1990), Larocheand Nioche (1994) for example.

Causality relations are very interesting becausethey are really meaningful. But this assumes that theproblem is already solved and that the set of pieces of information to be used in the causality chain have been identified. This is not the case in the field of strategic observation. In fact:

The chain is made up of only a few pieces of information (incomplete information) The pieces of information are not ordered: we

have a cause without its effects or an effect with-out its causes.

The causality relation may be considered as anideal situation that is probably unachievable andprobably inaccessible, especially under daily pressureand with limited means. We therefore have to find


8/1585



another solution which is why have chosen to con-sider more soft connections that are undoubtedlymore usable in this case.

This fits with the so-called influence link (Roos& Hall, 1980). In this case, information A is connectedto information B if A has influence on B, withoutconsidering A as the single cause of B and withoutrequiring A to be the direct and single cause of B. Thistype of link has been used many times (Lesca, 1989).

The Object To LinkWhen they are faced with inaccurate and ambigu-

ous data, and unable to make connections usingcausality or influence links, individuals may try tooppose them. This dialectic has been analyzed bytheorists in psychology. They state that people are better able to understand the opposition or difference between two concepts than each piece of informationtaken separately (Moles, 1990). Lee and Lai (1991)propose this type of relation seeing it as effective wayof drawing meaning from raw information. TheSIBYL software uses opposition relation to supportgroup decision making (Lee, 1990). In the BI field,where the process deals with signals that are forerun-ners of changes, the relation of opposition could beused to bring out inconsistencies or incoherence be-tween pieces of information that have been groupedtogether. It is critical to note that, whatever the type of link we choose to create meaning from disparatepieces of information, a difficult problem has to besolved: how to set up the choice of a link between twopieces of information?

The Confirm To LinkAccording to Hunt and Zartarian (1990), one of

the best ways of assessing the credibility of weak signals is to seek reciprocal confirmation. To illustratetheir idea, the authors propose the following situa-tion. When you learn, both from a client and asupplier, that one of your competitors is preparing tolaunch a new product, you can consider the piece of information as certain. Unfortunately, pieces of infor-mation handled by managers do not come fromvarious sources, especially data which are very new.In such cases, managers have to create potential links between pieces of information in order to decide

whether mutual confirmation is possible. The Con-firm to link allows managers to evaluate the credibil-ity and accuracy of pieces of information, and enablesthem to transform weak signals into more reliableinformation.

Skeleton and Set Up of the SuggestedMethod

We can now sum up our knowledge and incorpo-rate it into a method to create collective intelligenceon the environment of an organization (see Figure 2).

We could have chosen to instrumentalize themethod using a software support tool, but we prefer

the paper method to evaluate theoretical proposalsonly. We chose to avoid all interference arising fromthe use and the acceptance of IT. Furthermore, ourfield experiences show that, in the early stage of implementation, the use of IT tends to be perceived astoo rigid and constraining, especially in small andmedium size firms.

New weak signal

Is it

linked up with an existing group ?

noyes

Is itnecessary tocreate a new

group ?

no

Rejection of the information

yesFirst stepGroup enrichment

Criteria:regroupment by similarity and

proximity

First stepCreation of a

new group

Second stepCreation of links between group of informations

VISUAL SYNTHESIS

Criteria : causality link , object to link ,confirm to link, and hypothetical link

Hascreation produced

a trigger ?

no

Action

yes

noyes

Creation of a secondvisual synthesis

Action Setting informationson waiting

Hascreation produced

a trigger ?

Figure 2: Creative Collective IntelligenceProcess: A Conceptual Model


9/1586



Research ApproachTo some extent, our research methodology is simi-

lar to action research in the sense we have attemptedto produce results of practical value to organizationswith which we are allied while at the same timeadding to theoretical knowledge and gaining exper-tise. This approach can therefore be useful both topractitioners hoping to implement or improve BI intheir organization and to researchers seeking a betterunderstanding of BI.

Another important point in our approach is the building of tools as in an engineering approach. Thesetools are contributing to enrich a BI technologicalplatform or BI tool kit thus enabling the evolutionof BI. This approach gives practitioners the means totackle BI issues and enable researchers to observe theBI process in details and to collect data on the way BIis performed. This research can be compared to sys-temic action research or evolutive research as under-stood by Myer and Avison (2001).

Prerequisites for the Implementation of Management Engineering Approach

Before engaging in this kind of engineering work,

we need to address the following points: Nature of the problem : complex and poorly struc-tured, arising from field observations and neverpreviously addressed

Type of knowledge to be produced: the aim of thework should include the building of methods andtools to further BI performance. This knowledgeshould provide useful representations and in-sights into the BI process to support action andimprove conceptual and theoretical knowledge

Stages in Implementation of Management Engineering Research

The starting point is the existence of a gap be-tween empirical problems and scientific literature.The need for greater intelligibility regarding both theliterature and concrete situations has led us to designa theoretical model of the process. The engineeringaspect of the methodology consists in building amethod - i.e. a prototype of our model - in order to

implement it within organizations and enable us tomake relevant observations. The method, with itsuser guide, is then tested within organizations inorder to: 1) improve the situation from the practitio-ners standpoint and 2) collect data on what is satis-factory or unsatisfactory in the new method. Thiscontributes to an empirical validation of theoretical

assumptions and managerial beliefs. Figure 3 depictsa schematic representation of this management engi-neering approach.

Implementation within organizations lasts at leastfor a few months and focuses on the identification of a problematical situation, explaining it, and develop-ing a solution. This broad-based collaboration isakin to participatory action research (Whyte, 1991).

Criteria for Assessing ExperimentsThe results that need to be brought out to assess

the work achieved are at once diverse and comple-mentary as shown below:

Intelligibility of a complex process and ability togain a satisfactory image of it

Support for the understanding of a problematicalsituation

Comprehensibility and usability of theoretical con-cepts presented in the support method and articu-lation of disparate theoretical knowledge

Figure 3: Stages in ManagementEngineering Research

Problem identifiedwithin organizations

TheoreticalModel

Badly structured ornonexisting scientific knowledge

Operational Model building of a method

Implementationwithin organizations

External validity of the model

Internal validity of themodel and scientific contribution

Theoretical LoopPractical Loop


10/1587



Table 2Major Contributions of Our Research

Cognitive Process for the Exploitation ofStrategic Information

Business Intelligence Process

Conditions for Strategic InformationProcessing

Practicability of the Method

Utility of the Method

Managers have a tendency to process weak signals holistically Managers have tendency to regroup pieces of information by

subject The causality link is not suited to the process of strategic

information Managers take the reliability of strategic information into

account Linking pieces of information is useful for processing weak

signals but managers perceive this as a difficult task.

Our results confirm that BI is perceived as a complex process.Managers perceived it as difficult to implement and organize.

Our results confirm that BI is an iterative process with feedback loops. Analyses of strategic information must inducefirms into taking strategic decisions, but can also lead them tomodify the targeting stage.

Processing of weak signals must be done by experts Processing of weak signals is effective only if the firm has

formalized the targeting, tracking, selecting and circulation stages.

The method enables managers to process weak signals and tocreate visual synthesis. The proposed criteria seem to beadequate and easy to use.

The method enables firms to progress in the processing of weak signals and at the same time in the creation of visualsynthesis to support strategic decisions.

Expected Contributions Theoretical Contributions

Identification of hypotheses to allow the model to be generalized and the approach to be repro-duced.

Implementation of the Conceptual ModelFirst, it should be noted that the internal validity

of our conceptual model has been validated throughlaboratory experiments. After further refining, themodel was implemented within organizations.

Data SourcesFor this article, we worked with four medium-

sized companies. A necessary criterion to undertake

work with them was the perception by BI practitio-ners of difficulties in analyzing weak signals. Thesefour companies operate within turbulent environ-

ments such as telecommunications, microelectronics,and banking which explains their interest in BI. Eachcollaboration lasted six months on average.

Design for Data CollectionTo meet our objectives, data collection relies largely

on sessions calling on a collective learning process.The process has four stages (Davis & Olson, 1995):raising awareness, individual learning, recommenda-tions, and validation. Between each collective learn-


11/1588



ing session (once a month), people used the proposedmethod on an ongoing basis to assess its usability andusefulness. Their remarks were collected for analysisduring the next collective session. The method canthus be collectively refined and worked out in moredepth according to the specific features of each orga-nization.

Procedures and Means for Collecting andRecording Data

At least two researchers were present during thecollective session, one to run the support method andthe other to collect data through direct observations.A data collection grid was provided for the observer.This was in the form of a knowledge database (appro-priateness of tackled concepts, level of understanding by practitioners, articulation of concepts) and in-cluded observations on theoretical concepts presentedin the support method:

assessment of the model (relevance of the concep-tual model to empirical situations, identificationof strengths and weaknesses of the model)

assessment of the support tool and researchmethod (perceived satisfaction and utility, valid-ity of data collection using the tool, potentialimprovements, perceived completion of each stagein the research method)

Findings and Contributions

Findings: Creating Collective BusinessIntelligence

Our findings highlight the need for a method, theoverall satisfaction and usefulness of our prototype,and the ability to achieve meaningful and anticipa-tory representations of organizations environments.Figure 4 illustrates the concept of meaningful repre-sentation.

With regards to Figure 4, our research team wasinterested in IBM and its possible strategy regardingservice activities. We collected 11 potentially interest-ing weak signals in order to build a meaningfulrepresentation of IBMs services policy to gain betterunderstanding of its strategy. The central idea of therepresentation is in the form of a question: Is IBMmoving towards a service based strategy?

Two weak signals (IBM is disappointing its cus-tomers and IBM is reorienting its strategy) allowedus through a causality link to validate the idea thatIBM has indeed decided on a strategic shift towardsservices. We then tried to confirm the new policy bylooking to see whether IBM had given itself themeans to pursue this new policy.

The use of confirmation links enabled us to con-firm that IBM is profoundly motivated in favor of ashift towards services. For example, it has engaged aCEO for services (a CEO has been appointed to

service activities) and has tried to change its struc-ture (IBM intensifies the division between softwareand services).

Some contradictions have appeared like IBMspolicy of increasing its control over its subsidiariesand its moves to restructure itself as multiple compa-nies. Due to this contradiction, it was consequentlyappropriate to check the reliability of these two piecesof information. It was also necessary to eliminate thecontradiction by adding new information. The use of Figure 4: Example of a Meaningful

Representation of the Environment

IBM does want to monitor all its subsidiariesCollected signals

Explanations andcommentaries

Central hypothesis

Blindspots, actionsto undertake

KEYS

"Confirm to" links

"Object to" link

"Causality" link

IBM would bedividing itself intoseveral companies

Except services, whatabout other strategic

orientations (networks...)?

Will there be aspecialized entity

dedicated to services.Information to obtain

in urgence

IBM is reorientingits strategy

Is IBM movingtowards a services-

based strategy?

Means that are being implementedconfirm this new strategic orientation

Confirmation of their new strategy

IBM isreorganizing its

salesforce

A CEO has beenappointed to

services activities

Services supplying will be the subject of contract between

IBM and its customers

What more aboutthese contracts:

are customers satisfied?

IBM intensifies theseparation between

software and services

IBM is more and morecarefully listening to

its customers

IBM is disappointingits customers

A priority for IBM isto know its

customers' needs

Contradiction:Reliability ofthese signals?


12/1589



the hypothetical causality link brought new questionsto light, which were usefully informed by the contribution of new pieces of information or weak signals.

To conclude, the method made it possible toanswer part of our question by bringing in elementsof confirmation. However, a number of contradictionsappeared leading us to seek new information. Someanalyses and comments also call for further informa-tion.

Practical and Theoretical ContributionsFor easier comparison between expected and ac-

tual contributions, we have presented our main con-tributions in tabular form (see Table 2).

Detailed Results on the Utility andPracticality of the Method

Utility of the MethodOf all the practitioners who were interviewed,

only two reacted unfavorably to the method. The firstthought that the method was not useful for thecompany because its environment was not particu-larly complex and could easily be understood with-

out a specific method. The second practitioner criti-cized the contri bution of a method that does not startfrom the beginning of the BI process. He was bothered by the fact that the method contributes to the exploita-tion of information without explaining how it wascollected and circulated.

The observations from the other practitioners makeit possible to come to a favorable conclusion as to theacceptability of the method:

It implements a natural way of working. It pro-

poses, but in a more formalized way, an intuitiveway of working

The method is perceived as a factor of progress inthat it formalizes a complex process

It allows knowledge to be confronted and espe-cially offers real help in exploiting weak signals

It offers true solution to the lack of know-howamong practitioners

One of the practitioners stressed that the methodoffers a support tool for the organization andhelps to dynamize the whole BI process

Finally, presenting the information in the form of a visual synthesis seems to be useful as all theinformation can be shown on the same support,which make it easier to interpret

In short, the results obtained from the practitio-ners validate our production to some extent as well asthe assumptions made regarding acceptance by man-agers if a method for the exploitation of strategicinformation like weak signals were proposed to them.

Praticality of the Method

Stage 1: Regrouping Pieces of Information:Our experiments with companies make it possible

to validate the phase of regrouping pieces of informa-tion as well as the relevance of the suggested criteria,which seem to have been accepted quite naturally bythe practitioners. Some of them even found themperfectly obvious. Moreover, some companies ac-knowledged that they were already using these crite-

ria to exploit their information.Stage 2: Connecting Pieces of Information:

The managers thought that the method made thetask of connecting pieces of information easier. Inparticular, the written form of the suggested linksgained widespread approval among the practitioners,who thought that this helped to make fast compari-sons between pieces of information and to interpretrepresentations during subsequent consultations.

This stage also made it possible to validate infor-

mation and to reflect on the real relationships be-tween pieces of information. It helped to take theirambiguous nature into account.

The visual presentation of the information as wellas the written form of the links had the advantage of providing a visual synthesis that could be readilyunderstood, interpreted and communicated. The prac-titioners were attracted by the holistic reading ofrepresentations: they could quickly and easily recon-


13/1590



stitute and understand the reasoning from which thesynthesis was built upon.

Stage 3: Leaving a Trace of Reasoning:All the practitioners agreed on the importance

and usefulness of this stage. According to them, this isthis stage where the method demonstrates its greatestvalue. The method appears genuinely helpful in thatit formalizes what managers had already intuited andallows more in- depth reasoning.

The method thus appears to be very useful as itprovides practitioners with an effective method of operation. In conclusion, the major findings regard-ing to the use of the prototype are:

it is pragmatic and therefore easy to use it is a communication tool and concept which

fosters knowledge distribution and mutual en-richment through dialogue

it is an organizational tool and concept for theimplementation of the BI process particularly withregard to the daily exploitation of information

it is a training tool which shows how to solve theparadox of perceived information overload andlack of information

ConclusionDespite the need among practitioners for an ap-

propriate strategic information support method toexploit weak signals, only limited knowledge wasavailable. The exploratory nature of our research thusled us to develop a qualitative methodology whichhas produced practical and theoretical results that arequite encouraging. First, a complex process has beenmade more intelligible thanks to a conceptual modelintegrating both systemic and strategic dimensions.

The instrumentation represents an initial explorationinto what a suitable set of mechanisms for the cre-ation of collective intelligence might be. Future re-search could use our model as a framework for thedevelopment of information systems aimed at collec-tive knowledge creation. Secondly, articulating knowl-edge on BI not only improves the ability of practitio-ners to make decisions within turbulent environ-ments but also helps them to identify hypotheses thatare likely to improve the BI process.

ReferencesAnsoff H.I. (1975). Managing Strategic Surprise by

Response to Weak Signals, California Management Review 18(2): 21-33.

Barr P.S., Stimpert J.L and A.S. Huff. (1992).Cognitive Change, Strategic Action andOrganizational Renewal, Strategic Management Journal 13 (Special Summer Issue): 15-36.

Behling O., Gifford, W.E. and J.M. Tolliver. (1980).Effects of Grouping Information on DecisionMaking Under Risk, Decision Sciences 11(2): 272283.

Bougon M.G., Weick K.E. and D. Binkhorst. (1977).Cognition in Organizations: An Analysis of theUtrecht Jazz Orchestra, Administrative ScienceQuarterly 22(4): 608-639.

Conklin J. (1987). Hypertext: an Introduction andSurvey, IEEE Computer 20(9): 17-41.

Davis G.B. and M. Olson. (1995). ManagementInformation Systems: Conceptual Foundations,

Structure and Development . New York, NY:McGraw-Hill.

Feldman, M.S. and J.G. March. (1981). Informationin Organizations as Signal and Symbol,Administrative Science Quarterly , 26(2): 171-186.

Goldhar J.D. and Bragaw, L.K. and J.J. Schwarts.(1976). Information Flows, Management Stylesand Technological Innovation, IEEE Transactionson Engineering Management (23)1: 51-62.

Gorry G.A. and M.S. Scott-Morton. (1971). AFramework for Management InformationSystems, Sloan Management Review 13(1): 55-70.

Hogarth R.M. (1980). Judgement and Choice. ThePsychology of Decision. New York, NY: John Wileyand Sons.


14/1591



Hunt, C. and V. Zartarian. (1990). Le renseignementstratgique au service de votre entreprise:linformation pour gagner. Paris, FR: First.

Laroche, H. and J.P. Nioche. (1994). Lapprochecognitive de la stratgie dentreprise, RevueFranaise de Gestion (June-July-August): 64-78.

Lee J. (1990). SIBYL: A Tool for Managing GroupDesign Rationale, pp.79-92 in Proceedings of theConference on Computer Supported CooperativeWork, October 7-10, Los Angeles, CA. New York,NY: ACM Press.

Lee J, and K. Lai. (1991). Whats in DesignRationale? Human-Computer Interaction , 6 (3&4):251-280.

Lesca H. (1989). Information et adaptation delentreprise. Paris, FR: Masson.

Lesca H. (1992). Le problme crucial de la veillestratgique: la construction du puzzle. RevueAnales des Mines (June): 67-71.

Lesca H. (1994). Veille stratgique pour lemanagement stratgique: tat de la question etaxes de recherche, in Economie et Socits. SrieSciences de Gestion SG (20)5: 31-50.

McKenney, J.L. and P.G.W. Keen. (1974). HowManagers Minds Work, Harvard BusinessReview 52(3): 79-90.

Miller G. (1956). The Magical Number Seven, Plusor Minus Two: Some Limits on Our Capacity forProcessing Information, Psychological Review63(2): 81-97.

Mintzberg, H. (1976). Planning on the Left Side andManaging on the Right, Harvard Business Review5(4): 49-58.

Moles A. (1990). Abaques de regnier, pp. 225-228in Les Sciences de Limprcis. Paris, FR: Editionsdu Seuil.

Myer M.D. and D. Avison [eds.] (2001). QualitativeResearch in Information Systems: A Reader. London,UK: Sage Publications

Narayanan V.K. and L. Fahey. (1990). Evolution of Revealed Causal Maps During Decline: A CaseStudy of Admiral, pp.109 -133 in Huff, A.S.[Ed.] Mapping Strategic Thought . Chichester, UK: John Wiley and Sons.

Roos L.L. Jr. and R.I. Hall. (1980). InfluenceDiagrams and Organizational Power,Administrative Science Quarterly 25(1): 57-71.

Rowe F. and A. Ziti. (2000). Cognition individuelleet systmes dinformation, Systme dinformationet de management (SIM) 4(5): 3-20.

Simon H.A. (1982). Models of BoundedRationality, Cambridge, MA: MIT Press.

Taggart W. and D. Robey. (1981). Minds andManagers: On the Dual Nature of HumanInformation Processing and Management,Academy of Management Review 6(2): 187-195.

Weber C.E. (1984). Strategic Thinking: Dealing withUncertainty, Long Range Planning , 17(5): p 60-70

Whyte W.F. [ed.] (1991). Participatory Action Research.New York, NY: Sage Publications.

Selected Bibliography

Books

Demory B. ( 1983).La Crativit en 50 Questions.Paris, FR: Chotard et Associs.

Galliers R.D. (1996). Choosing Information SystemsResearch Approaches, pp. 144-162 in R. Galliers[ed.]

Information Systems Research: Issues, Methods andPractical Guidelines. Oxford, UK: BlackwellScientific Publications.


15/15


Huff A.S. [ed]. (1990). Mapping Strategic Thought .New York, NY: John Wiley and Sons.

Lesca H. (1986).Systme dinformation pour lemanagement stratgique de lentreprise: lentrepriseintelligente. Paris, FR: MacGraw Hill.

Mintzberg H. (1990). Le management. voyage aucentre des organisations. Paris, FR: EditionsdOrganisation.

Richard J.F. (1990). Les activits mentales:comprendre, raisonner, trouver des solutions.Paris, FR: Armand Colin.

Book Chapters

Tricker R.I. (1992). The Management of Organizational Knowledge, pp.14-27 in R.Galliers [Ed]. Information Systems Research:Issues, Methods and Practical Guidelines. Oxford,UK: Blackwell Scientific Publications Ltd.

Academic Journal Articles

Bariff M.L. and E.J. Lusk. (1977). Cognitive andPersonality Tests for the Design of ManagementInformation Systems, Management Science 23(8):820-829.

Couger J.D., Higgins L.F., and S.C. McIntyre. (1993).(Un)structured Creativity in InformationSystems Organizations, MIS Quarterly 17(4):375-397.

Getz I. (1994). Systme dinformation: lapport dela psychologie cognitive, Revue Franaise deGestion 99(June, July, August): 92-108.

Hurst D.K., Rush J.C. and R.E. White. (1989). TopManagement Teams and OrganizationalRenewal, Strategic Management Journal10(Summer Special Issue) 87-105.

Meyer A.D. (1991). Visual Data in OrganizationalResearch, Organization Science 2(2): 218-236.

Mintzberg H. (1972). The Myth of MIS, California Management Review 15(1): 92-97.

Mintzberg H. (1975). The Managers Job: Folkloreand Fact, Harvard Business Review 53(4): 49-61.

Schwenk C.R. (1984). Cognitive SimplificationProcesses in Strategic Decision Making,Strategic Management Journal 5(2): 111-128.

Sullivan C.H. Jr. and C.E. Yates (1988). Reasoning by Analogy: A Tool for Business Planning, Sloan Management Review 29(3): 55-60.

Susman G.I. and R.D. Evered. (1978). AnAssessment of the Scientific Merits of ActionResearch, Administrative Science Quarterly 23(4):582-603.

Yadav S.B. and D. Khazanchi. (1992). SubjectiveUnderstanding in Strategic Decision Making,Decision Support Systems 8(1): 55-71.

Westley F. and H. Mintzberg. (1989). VisionaryLeadership and Strategic Management, Strategic Management Journal 10(Summer): 17-32.

Date post:	04-Jun-2018
Category:	Documents
Upload:	cp1y-scribd9713
View:	215 times
Download:	0 times

2003-Blanco-caron-lesca-scipDeveloping Capabilities to Create Collective Intelligence

Documents